In the past ten years metastatic prostate cancer killed more than 400,000 American men and today remains the second most common cause of cancer death in U.S. men in 20001. The molecular basis of prostate cancer metastasis is poorly understood and has been rarely studied in human metastatic prostate cancer cells derived directly from patients with this disease. Moreover, the molecular basis for the increased incidence and poorer prognosis for prostate cancer in African-Americans remains uncertain. We hypothesize that alterations of upstream and downstream effectors of p53 gene function, and of the p53 gene itself, are important contributors to prostate cancer progression. This hypothesis will be tested through two specific aims: Specific Aim 1 will examine critical members of the p53 pathway (p53, Caveolin-l, p14, p21, MDM2, and ATM) in a series of 143 androgen-dependent and androgen-independent bone, lymph node, and visceral metastases from 81 patients. A combination of molecular analysis techniques will determine 1) whether uniform patterns of p53 pathway alteration occur in samples from multiple metastatic sites from individual patients and 2) whether specific patterns of alteration are associated with specific modes of dissemination of prostate cancer (e.g., bone only versus bone plus visceral sites). The most informative molecular techniques identified will be used to examine prostate cancer tissue removed at the time of initial prostate cancer diagnosis from a subset of 30 of the 81 total patients to provide insight into the timing of alteration of this pathway. Additional analysis will compare comparative genomic hybridization data to the p53 pathway data in parallel specimens to identify gains or losses that are associated with specific defects in the p53 pathway. Specific Aim 2 will test the hypothesis, based on recently published data2'3, that variations in the pattern of alteration of the p53 pathway may account for increased aggressiveness of prostate cancer in African-Americans. Results from a subset of 31 androgen-dependent and androgen-independent metastatic prostate cancers from African-Americans will be compared to samples from 50 White and Hispanic patients in this analysis. The results of these studies will 1) have a direct impact on future selection of gene therapy and other therapies offered to patients with prostate cancer, 2) will help evaluate the relevance of tumor models in which p53 inactivation is a critical factor, and 3) will help identify genomic data elements needed to define prostate cancer phenotypes.